French as a foreign language and the Common European Framework of Reference for languages

Author: James Milton


This paper considers the position which British GCSE and 'A' level are given in the Common European Framework (CEFR). The vocabulary sizes of learners taking these exams are considered in relation to the vocabulary information and wordlist sizes included in the CEFR documentation. The vocabulary knowledge of learners appears small, very small, in comparison to the levels anticipated by the Framework, and very small compared to learners of other languages at the same levels. Learners in Britain appear to lack the vocabulary knowledge necessary to carry out the skills indicated for the levels they are expected to attain.

This article was added to our website on 20/02/07 at which time all links were checked. However, we cannot guarantee that the links are still valid.

Table of contents

Conference 2006

This paper was originally presented at our conference: Crossing frontiers: languages and the international dimension, 6-7 July 2006. Download print version: this paper is also available as a pdf (121Kb)


This paper considers the place which our foreign language exams are given in the Common European Framework of Reference for Languages (CEFR). It is intended, also, to consider where other exams are placed and in particular the position of EFL exams but the focus of the paper is intended to be the French foreign language exams we use in the state system in Britain. The CEFR and the placement of exams within it will be tackled from the point of view of learners' vocabulary knowledge. It is worth recalling, at the outset, what the CEFR is for. It is intended to be a practical tool for setting clear standards to be attained at successive stages of learning and for evaluating outcomes in an internationally comparable manner. It is intended to aid the setting clear standards in an internationally comparable manner, to aid clear certification. This in turn will make our standards and awards clear and comparable to users: to students, to parents, and to employers. Also, where we teach for exams placed within this framework, is should set clear goals and standards to work towards. (The Council of Europe, 2003). It is intended to address the question of whether the placement of British French exams in the Framework is appropriate.

The CEFR and exam placements

Foreign language teaching has for many years worked to a three tier hierarchy of level and achievement: elementary, intermediate and advanced. The CEFR takes that hierarchy and extends and renames the elements within it. The three tiers are renamed Basic, Independent and Proficient. These three tiers are each sub-divided into two giving a six level framework as shown in Table 1.

Table 1: The CEFR hierarchy of levels
ProficientC1Effective Operational Proficiency

To aid placement of exams, and learners, within this framework, there is a substantial document describing and characterising the levels (The Council of Europe, 2003). These are often condensed to Can-Do descriptors whose intention is to be sufficiently general and inclusive to allow almost any course, exam, book or learner to be placed within it. In the latest form of the CEFR (The Council of Europe, 2003) these descriptors are skills based. Examining bodies, starting with the EFL world, have placed their exams within this system. Large examining bodies such as the Cambridge Syndicate have exams at each of the six levels. Here in Britain, the Department for Education and Skills do not endorse so many formal foreign language exams but have placed the major ones we use in schools, GCSE and 'A' level, within the system. But there is, or at least there should be, a higher level foreign language equivalent in Britain in the form of the foreign language degrees which, I presume, should be at the top end of the scale. If we, the universities, turn out users who are not at mastery level in French then we should be worried. In Table 2 the Cambridge English and French foreign language exams at each level are placed alongside each other. The question I am addressing in this paper is whether these placements appropriate and whether exams which should be equivalent in level really are so? This will question whether the CEFR is, as yet, the tool for international comparison it seeks to be.

Table 2: English and French foreign language qualification placements in the CEFR
CEFR levelCambridge ESOLFrench in UK
A1Starters Etc. 
B2FCE'A' Level
C2CPEBA in French?

The nature of the framework descriptors, and the nature of language itself, generally makes comparison of level between languages extremely hard. There are three reasons for trying to compare language levels using vocabulary size and vocabulary knowledge as an indicator. One is that the original work on the framework included wordlists so we do have some idea as to what vocabulary knowledge was expected of learners at some of the levels. A second is that vocabulary size and coverage are strongly connected so vocabulary size ought to be a good general indicator of more general knowledge and performance. Research bears out this idea (for example, Milton, 2006a) and vocabulary size does indeed appear a good general indicator of foreign language ability. The third reason is that vocabulary is countable: it is possible to put a figure to levels of knowledge and achievement. This fact allows comparisons between languages to be made, so it possible to compare knowledge between English and French as foreign languages, for example, in a way which is not usually possible.

The CEFR and vocabulary levels

The original Threshold level materials (Van Ek and Trim, 1990; Coste et al 1987) contained wordlists for French, and other languages at B1 level. The French list contains just under 2000 words, and the EFL list just over 2000 words, for example. The A2, Waystage materials for French and English (for example, Van Ek, 1990) both contain wordlists of about 1000 words. You might expect learners taking French foreign language exams at these levels to have vocabularies of something like these sizes.

In both English and French the most frequent 2000 words, and overwhelmingly the most frequent words in a language are learned earliest (Milton, forthcoming) give about 80% coverage of normal text. This is a very interesting and important figure because it marks the level at which learners appear to progress from understanding almost nothing they hear or read, except in the most limited and contrived of circumstances, to having passages of clarity and being able to grasp the gist of a conversation or a reading passage. But, in both languages, to add sufficient vocabulary to understand the remaining 20%, and therefore understand all of a text, requires massively more vocabulary. Learners do not have anything like full comprehension of a text until they have at least 95% or 98% coverage of a text and that may require 6000 to 8000 words.

In principle, then, we might expect learners of both English and French to have vocabularies of roughly the same size when they take exams at the same CEFR level. But what vocabulary levels the real learners know when, in reality, they take exams at these levels? Vocabulary size tests, in both languages, have been applied to learners across the whole range of language ability and the results of these investigations allow us to begin to answer this question. Figure 1 presents the mean scores for learners of French as a foreign language in school and university, at the end of each year of study up to degree level.

Figure 1: Vocabulary growth in French in a British school and university (Milton, 2006b)
Figure 1: Vocabulary growth in French in a British school and university (Milton, 2006b)

What appears to occur is that there is modest vocabulary growth at the outset of learning in school until year 5 when a portion of the learners are able to drop the subject and you get a jump in the mean. GCSE learners, at level A2 or B1 therefore, know on average about 850 words. Above this level presumably only the best learners select languages at 'A' level and they appear to add 500 or so words per year, but a proportion of the jump in level between GCSE and 'A' level is likely to be the effect of lower level learners dropping out of the group learning French. 'A' level learners, at B2 level, therefore, know about1920 words of French on average. The figures reveal that vocabulary size is an excellent predictor of the grade that 'A' level students obtain. Other studies suggest that these figures are probably typical of most schools and learners throughout Britain. Milton and Meara's (1998) study of school learners' French vocabulary knowledge in south Wales provide very similar figures for the levels that were tested. A recent study by Annabelle David of four schools in Newcastle (personal correspondence) also gives similarly sized estimates.

Another 500 words a year are added to the mean while learners are in university although, again, this may be partly accounted for by the drop in the numbers studying the subject after 'A' level. Except, it appears, in the year abroad. You might expect to see considerable gain in learners' vocabulary sizes during a period of study in the country where the language is spoken. Certainly, the overseas students learning English in UK have been demonstrated to add about 1500 words on average in an academic year, October to May during study at a British university (Milton and Meara, 1995). No data was available for this group directly and the year abroad is omitted from the graph, but no great progress is visible in the British learners who study in France from the evidence of learners' final year vocabulary size scores. Just possibly learners are making progress in other ways, and the learners themselves suggest that they feel the progress they make on the year abroad disappears quickly on their return. However, equally likely is the possibility that many learners live in an expatriate English speaking community whilst on the year abroad and do not have the amount or quality of input which would be required for significant progress. The number of learners this area of study is based on is small but this observation deserves much closer attention. Whatever the nature of progress during the year abroad, graduates complete their studies, on average, with a score of about 3300 words on this test. This would seem to represent a fairly modest level of knowledge and far short of native-like knowledge and fluency.

If these results are compared with the levels described by the Framework materials themselves, there is a big shortfall. GCSE Lower looks like it might, just might, be pitched at about the right level but GCSE Higher and 'A' level fall well short in vocabulary knowledge of the levels which were expected. With vocabulary sizes as limited as this the shortfall is highly significant and suggests that the learners must also be very limited in their communicative skill and ability. GCSE learners fall well short of the B1 Threshold level of 2000 words (and 80% coverage of normal text) and it would appear they are well short of the level needed to be independent users. 'A' level students, with under 2000 words on average, look like they are just hitting the vocabulary levels needed for gist understanding which would place them at B1 rather than B2 level. My observation is that this is about right; our university entrants struggle to hold a conversation on anything but the most predictable and limited of topics, and certainly could not follow a lecture or a seminar in French. French graduates, with about 3300 words on average, appear well short of complete mastery and again this calculation appears to match the observation of students in my own university. We no longer routinely lecture through the medium of French, we cut down the number of books students read, to enable them to cope, and much of the work in done in translation. Graduates cannot perform in all aspects of language in an educated native-like fashion.

If the results are little disappointing in French and for the UK exam system, how to other systems fare? Comparison across languages may not always be possible since it is feasible that differences in the structure of language would make it possible to carry out more language functions with fewer vocabulary resources in some languages than in others. However, there are probably good reasons for thinking that French and English vocabularies required for the various skills levels should be the same. Vocabulary in both languages is structured very similarly. While they may inflect and derive words slightly differently there is nothing like the kind of difference which might be seen if these two languages were contrasted with, say Finnish or Hungarian. Much vocabulary is actually cognate. Coverage and vocabulary size appears to work very similarly, and the word lists derives for the Threshold and Waystage levels are very similar in size. It would be a surprise if learners at the same level of ability and performance in these two languages had radically different vocabulary sizes.

Vocabulary growth figures are available for EFL learners in Greece who take the Cambridge suite of exams up to C2 level, and for Hungarian learners whose Maturity exams are pitched at A2, B1 and B2 levels. The figures are very similar and are shown in Table 3.

Table 3: Vocabulary knowledge at various CEFR levels in Britain, Greece and Hungary
CEF levelWordlist sizeFrenchEFL GreeceEFL Hungary
B2 192035003900
C2 33004500 

The results in the UK system fall short of the expected levels while the EFL learners exceed the expected levels by a similar margin. There is an enormous difference between the vocabulary sizes of British learners of French and European learners of EFL, at every level of the Framework where results have been obtained. There is no absolute right and wrong here. The vocabulary sizes I have placed in the Framework are not absolutes and an exact correspondence of vocabulary sizes in the two languages should not be expected. But the difference is very large and I think it should be a concern that in Britain we seem to have placed our standards at the lowest end of the whole Framework system. Scores from UK students appear low compared to the contents of the Framework itself and low compared to other students at the same levels.


At the outset of the paper the question was posed whether the placement of the UK French exams in the CEFR was appropriate and the answer is that probably they are not. It seems to me that this is a great pity. It reinforces the stereotype that the British are bad at languages: what users of the system expect these learners to be able to do, they cannot do. Because the UK system is quite different from the predominant one of the continent, and setting such visibly lower standards, it brings the system into disrepute. It fails to give the learners credit for the standards they have achieved, only discredit for not achieving the anticipated standards. Given the time available for teaching in Schools, achievement levels are very reasonable but we, within the British system, are making claims for them which are not merited. The time available in university for learning, particularly with the year abroad, is much greater but achievement appears less impressive. Perhaps it is asking too much of today's students entering university, with a comparatively low level of French, to take advantage of study abroad. The overseas students who come to UK are, overall, at a much higher language level when they arrive. Perhaps a university is not the best place to send our learners of French if their foreign language is to develop.


Coste, D., Courtillon, J., Ferenczi, V., Martins-Baltar, M. and Papo, E. (1987) Un Niveau Seuil. Paris; Editions Didier

Milton, J. (2006a) "X-Lex: the Swansea Vocabulary Levels Test." In Coombe, Davidson, and Lloyd Proceedings of the 7th and 8th Current Trends in English Language Testing (CTELT) Conference vol 4. UAE; TESOL Arabia, pp 29-39.

Milton, J. (2006b) "Language Lite: Learning French vocabulary in school." In Journal of French Language Studies 16(2006) pp. 187-205

Milton, J. (forthcoming). "Lexical profiles, learning styles and the construct validity of lexical size tests" in Daller, H., J. Milton, and J. Treffers-Daller, Modelling and Assessing Vocabulary Knowledge. Cambridge; CUP.

Milton, J. and Meara, P. (1998). "Are the British really bad at learning foreign languages?" Language Learning Journal 18 (December 1998) pp 68-76.

Milton, J. and P. Meara (1995) "How periods abroad affect vocabulary growth in a foreign language" ITL Review of Applied Linguistics 107-108 April 1995, pp 17-34

The Council of Europe (2003) Common Framework of Reference for Languages.Cambridge; CUP.

Van Ek, J.A. and Trim, J.L.M. (1990) Threshold 1990. Strasbourg: Council of Europe Publishing.

Van Ek, J.A. (1990) Waystage English. London; Pergamon Press